[r2003-12-09 00:26:42 by afrisch] Two new papers

Original author: afrisch
Date: 2003-12-09 00:26:42+00:00

web/index.xml

@@ -53,6 +53,19 @@ the content by <a href="examples.html#site">the following CDuce program</a>.
 </left>
 
 <box title="Latest News" link="news">
+<section title="2003, December 8th, Monday: New addition to papers.">
+<p>
+A paper on the optimized implementation of pattern matching
+in CDuce is now available online in the <a href="papers.html">papers page</a>:
+<i>"Regular tree language recognition with static information"</i>.
+Another paper, <i>"Greedy regular expression matching"</i>
+is a preliminary step in an alternative implementation technique
+(non-uniform representation of values). The two papers will be
+presented in the PLAN-X 2004 workshop in January (Venice).
+</p>
+<br/>
+</section>
+
 <section title="2003, October 26th, Sunday: New addition to papers.">
 <p>
 If you want a simple and basic introduction to the theoretical foundations

web/papers.xml

@@ -166,6 +166,64 @@ surcharg
 </ul>
 </section>
 
+<section title="Implementation algorithms">
+
+<ul>
+<li>
+<paper file="papers/reg.pdf">
+ <title>Regular tree language recognition with static information</title>
+ <author>A. Frisch</author>
+ <comment> In <i>PLAN-X 2004</i>. </comment>
+
+ <abstract><p>
+  This paper presents our compilation strategy to produce
+efficient code for pattern matching in the CDuce compiler, taking into
+account static information provided by the type system.  Indeed, this
+information allows in many cases to compute the result (that is, to
+decide which branch to consider) by looking only at a small fragment
+of the tree.  Formally, we introduce a new kind of deterministic tree
+automata that can efficiently recognize regular tree languages with
+static information about the trees and we propose a compilation
+algorithm to produce these automata.
+ </p>
+ </abstract>
+</paper>
+</li>
+</ul>
+
+<ul>
+<li>
+<paper file="papers/greedy.pdf">
+ <title>Greedy regular expression matching</title>
+ <author>A. Frisch</author>
+ <author>L. Cardelli</author>
+ <comment> In <i>PLAN-X 2004</i>. </comment>
+
+ <abstract><p>This paper studies the problem of matching sequences against regular
+expressions in order to produce structured values.  More specifically,
+we formalize in an abstract way a greedy disambiguation policy and
+propose efficient matching algorithms.  We also formalize and address 
+a folklore problem of non-termination in naive 
+implementations of the greedy semantics.</p>
+
+<p>
+Regular expression types and patterns have been introduced in the
+setting of XML-oriented functional languages.  Traditionnaly, all the XML
+values and sequences share a common uniform runtime representation.
+Our work suggests an alternative implementation technique, where
+regular expression types define not only a set of abstract
+flat sequences, but also a custom structured representation
+for such values. This paves the way to a variety of language designs
+and implementations to integrate XML structural types 
+in existing languages (class-based OO languages, imperative features,
+constrained runtime environment, ..).
+ </p></abstract>
+</paper>
+</li>
+</ul>
+
+</section>
+
 <section title="Security">
 
 <ul>